Qii.AI’s Platform for Drone Inspection Data: Label, Train, and Manage

Label, Train and Manage With Qii.AI Platform for Drone Inspection Data

By: Dawn Zoldi

For industrial inspections, data is king. Drones, equipped with a variety of sensors, can gather an overwhelming amount of it. Businesses don’t just need any data. They need the right data at the right time. Qii.AI, a Toronto, Ontario Canadian startup, helps companies hone in on what is important and relevant to them, and then to sort through it all.

Qii.AI’s artificial intelligence (AI)-machine learning (ML)-enabled platform coalesces mission-specific drone visual and digital inspection data into actionable intelligence for decision makers by providing targeted reports and 3D digital twins and disseminating them to all relevant stakeholders in near real time.

“Our mantra at Qii.AI is ‘Label, Train, and Manage’,” explained the company’s CEO Michael H. Cohen, a 20-year commercial pilot and former Boeing airline Captain, who initially started the company in 2014 to perform building inspections in Canada.

This initial use case employed drones with thermal cameras to image buildings at night for energy loss. Cohen’s team was the first to obtain waivers for tablet-driven autonomous night flights in both Canada and the United States.

“We were collecting about three to five gigabytes for each building,” Cohen recalled. He and his partners soon realized that they had an internal business need for data management and data analytics in these remote digital inspections – and that their clients did too.

“If you have 5000 photographs, the user just wants to see the best ones. How do you sort through those? You create ways in which algorithms can look at camera position, the lighting and various attributes to narrow those 5000 images down to the 20 that provide a holistic, 360 degree view of the asset,” he said.

Those algorithms require accurate comparable data-sets to narrow outputs to relevant images. Uniquely, Qii.AI engages their customers in the AI process, putting AI at the edge and allowing users to label the data. Cohen and his team call this “user-defined” labeling.

For example, in construction, the user can tell the computer exactly what is a crack and what isn’t a crack. “We learned from the trenches,” said Cohen. “Qiii.AI initially tried to pay third party corrosion experts, crack experts, inspectors, engineers to do this labeling. If you do find them, they’re prohibitively expensive to hire for creating, annotating, segmenting and labeling corrosion and crack data for concrete.”

On the other hand, business owners know their business best and speak their own language. “We love the user to provide the input because it gives them the control to produce mission-specific quality data,” Cohen added.

Qii.AI is the only enterprise AI platform that combines drone inspection software with a user-defined AI labeling tool.

The company then sends the data off for training. In ML, the more quality data that one provides, the better the predictive outcomes will be for future anomaly detection operations. Qii.AI’s smart system learns and builds upon itself, continuously improving the speed and quality of inspections.

Cloud-based software also enables users to collaboratively and remotely review the data through reports that stitch 2D photos into 3D digital twins. Digital twins help businesses predict the future condition of assets and equipment by showing how assets have changed overtime. This allows for repair scheduling at the optimal time, reduces operations costs and prevents catastrophic failures.

Based on internal employee brainstorming and consistent with the company’s “pushing the limits” spirit, it recently injected augmented reality (AR) into this backend process. It allows users to don AR goggles and immerse themselves into the 3D digital environment. AR allows management, an engineering company or an insurance company, often in different locations around the globe, far from the drone pilot and onsite inspection crew, to not only review the asset set for their own purposes in near real time, but to now also virtually transport themselves into the asset first-hand.

In short, Qii.AI automates drone inspections to complete them faster and with greater accuracy. User-defined AI gives users complete control over drone inspection data, from how images are  segmented and labeled, to the content included in reports. Digital twins provide a complete view of assets. AR takes the users into the review process.

Cohen said the company expects to make several exciting announcements soon, so stay tuned for more from Qii.AI.

Watch the recent Dawn of Drones Podcast with Qii.AI CEO Michael Cohen, part of the Carahsoft Partner Series.

Dawn M.K. Zoldi (Colonel, USAF, Retired) is a licensed attorney with 28 years of combined active duty military and federal civil service to the Department of the Air Force. She is an intIernationally recognized expert on unmanned aircraft system law and policy, a columnist for several magazines,recipient of the Woman to Watch in UAS (Leadership) Award 2019, President and CEO of UAS Colorado and the CEO of P3 Tech Consulting LLC. For more information, visit her website at: https://www.p3techconsulting.com.